Title: Research on collaborative recommendation of multidimensional sparse data based on personalised directional information fusion algorithm
Authors: Yi Huang
Addresses: School of software, Hunan Vocational College of Science and Technology, Changsha City of Hunan Province, 410004, China
Abstract: In order to solve the problem of poor precision of recommendation results in the traditional collaborative recommendation method for multidimensional sparse data, a collaborative recommendation method for multidimensional sparse data based on personalised directional information fusion algorithm was proposed. The cosine similarity of the data vector was calculated and modified, and the score prediction set was established. On the basis of the predicted value and variable quantity value, the median, mean and model were used to populate and deconstruct the standard representative data, construct the multiple score matrix, solve the data location problem of sparse matrix, and realise the data collaborative recommendation. The experimental results show that the average error of the research method is about 0.03, lower than the traditional method, which proves that the method can effectively improve the accuracy of data recommendation.
Keywords: directional information; fusion algorithm; data quantity; recommendation; multidimensional sparse data; predicted value; variable quantity value.
International Journal of Autonomous and Adaptive Communications Systems, 2020 Vol.13 No.4, pp.311 - 328
Received: 10 Oct 2019
Accepted: 21 Nov 2019
Published online: 12 Jan 2021 *